The overall long-term objective of this program project is to identify novel dietary, hormonal and genetic determinants of breast (Project 1), colorectal (Project 2), and ovarian (Project 3) cancer risk in women, with the ultimate aim of finding means for prevention and improved survival. The combination of questionnaire-derived data and biomarkers, coupled with the long-term follow-up, affords the opportunity to further understand of the time course, as well as mechanisms of cancer development. To achieve these objectives, we will relate a) Prospectively collected data on diet, weight gain, physical activity, analgesic use, and other behaviors; b) Nutrient and hormone levels in prospectively collected blood; and c) Genotypic information from archived DNA and tissue blocks; to incidence of breast, colorectal, and ovarian cancers. In addition, a fourth project will refine methods for polychotomous logistic regression and, using both a previous diet validation study and a new diet validation study, will address methodologic issues involving repeated measures of nutrients in relation to cancer risk. This project will also address causal inference and incidence modeling, as well as statistical issues in haplotype estimation and haplotype-environment interactions. Additional follow-up is required to identify adequate numbers of cases to address hypotheses outlined in each of the projects. This program project is based on the Nurses' Health Study cohort, comprising 121,700 women who were 30 to 55 years of age when enrolled in 1976. The program project serves as the central resource for updating questionnaire data and confirming incident diseases necessary for the many related grants addressing incidence of cancer and other major chronic diseases. Thus, through funding of Cores A and B in the program project, we maintain follow-up of the cohort and timely confirmation of incident cancers and deaths. Core C conducts specimen handling and genetic analysis and Core D provides leadership and data analysis. The program project also helps to support our repository of biological samples. With this renewal application, we have an institutional commitment of $386,000 to upgrade our freezer facility. We address three major themes across the projects. Each project will be providing either an initial assessment (ovarian cancer) or more in depth assessment (breast and colon cancers) of the role of one carbon metabolism and gene hypermethylation in carcinogenesis. Further, modeling these complex exposures will be a focus of Project 4, hence optimizing the analyses being conducted in the other three projects. The role of energy balance throughout life (as assessed through body size, physical activity, and IGF) is another focus of the three projects. The role of inflammatory markers in disease is an area of focus in both the colon and ovarian projects. Project 4 develops statistical approaches to data analysis that apply directly to Projects 1, 2, and 3 for evaluation of nutrient-cancer associations when using repeated measures of diet. Methodologic work on polychotomous logistic regression is applicable to each project as sufficient numbers of cases allow for evaluation of hypotheses relating to subsets of cancer endpoints (e.g., ductal vs., lobular breast cancer; mucinous vs. nonmucinous ovarian cancer; total cancer mortality). Furthermore, the development of methods for haplotype estimation and the evaluation of haplotype-environment interactions are applicable across each of the projects as we evaluate potential gene-environment interactions. Strong scientific synergy permeates the projects. All projects are based on the same infrastructure of data collection, management, disease follow-up, and data analysis. The investigators, although focusing on their own areas, interact closely, which allows cross-fertilization of knowledge, ideas and approaches, and results in considerable productivity and efficiency.